Application of new <i>Q</i>-<i>W</i> relation to study latent heat fluxes over Indian Ocean using Seasat SMMR data

IndianJ~fRadio ~Physics Vol~ember 199'f, pp.375-379

<"~

\,

J-\

~ Application of new Q- Wrelation to study latent ~t fJ.ux~over Indian

Ocean using S_ea_s_at'SMMRdata _{---- M}

:c..YovJ~"C 1 '" Q..\C^{_I ' ~}

__---'v~..~

'#"'_1J .. 4~.;P~

~N.:pa:utam,~.Basu, ~ M Gairola&.PCPandey \

C.

Meteorologyand OceanographyDivision,Space Applications Centre (ISRO),Ahmedabad 380~

Reeefved,:1~y1994;revisedreceived168eRt~1994

'The MONEX-79 ship observations have been used to compute latent heat flux (called direct fneiliod) over the Indian Ocean region. Latent heat fluxes have also been calculat using the above m~tioned ship observations except the direct measured values of surface level humidity (called indirect method) which have been derived using the relation proposed by Gautam et aL [Bound-Layer Meteoro/

(Netherlands), 60 (1992) 179]. The correlation coefficient between latent heat fluxes obtained by these two methods is 0.9, suggesting that this Q-W relation can be used for latent heat flux computations.

This methodology has also been extended to Seasat-SMMR data. Using the same data set, the latent heat flux .variabiltiesfor deep depression and normal cases over the Bay of Bengal have been studied to bring out the finer details of the ocean-atmosphere interaction process at the boundary

m.

^{This study}

demonstrates the application of satellite data for the estimation of heat fluxes on instantaneous basis for operational purposes in numeric~ models.,

I

^{il.- \ , •}

//.J-f ~~a'

"I Introduction precipitable water (W) (Ref. 4) for the estimation , The atmosphere and ocean form a strongly of

lEE

Gairola et aL5 applied Liu's approach6 for coupled system, especially in tropics. Short-term the estimation of LHF over the Indian Ocean us- climate changes are believed to be strongly influ- ing 3-day averaged Seasat-SMMR data. This study enced by large-scale ocean-atmosphere interaction was based on the suggestion of Hsu and Blan- through exchanges of momentum, heat and water. chard7 that monthly mean global Q- W relation of The latent heat supply is. a crucial driving force Liu can be used for instantaneous data sets. Re- for the tropical atmospheric circulations. So, mea- cently Gautam et aL⁸ examined the applicability of surements of latent heat fluxes {hereinafter LHF) Liu's global ^Q-W relation for instantaneous. data over scale of ocean basins are fundamental in the set using MONEX-79 data and found that Liu's understanding of interannual or long-term climatic global Q- W relation (Ref. 4) is not valid, over the fluctuations. Due to vast ~~c-data-s~~ re- Indian Ocean. They also derived a new relation gions, it is difficult to assess the in situ processes between Q and W using instantaneous MONEX- which allow study of long-term variability of eva- 79 data set. In the present paper, the LHF using poration and heat fluxes. SateIfites, however, pro- ^{Q- W}relation of Gautam ^{et aL8}and its distribution vide needed coverage to study long-term variabil- over the Bay of Bengal in deep depression (13-17 it}'. The passive microwave radiometers onboard August) and normal (7-11 August) cases have been Seasat/Nimbus-7 .stellit~ have already shown studied.

that several geophysical variables such as sea sur...:..,

face tempernhlre. (SST), wind spee<i (WS) and -2 Data base

£ectpltable water ^(W), etc. can be measured with The MONEX-79 data have been used for com- usefUl accuracies needed for air-sea interaction paring LHF calculated by direct method, i.e. by

studies1•2• using ship observations of SST, surface level bu-

Liu3 has shown the importance of satellite pas- midity, and WS, and LHF calculated by indirect sive microwave datil for the estimation of LHF on method, i.e. by using SST, WS and W from ship monthly scale using Seasat-SMMR data. Liu4 fur- observations and surface level humidity from ^{Q- W} ther .extended the work and estimated moisture equation given by Gautam et at.8

and latent heat flux variabilities using Nimbus-7 The SMMR (Scanning Multichannel Microwave data. He has applied' the global relation· between RadiometW on ocean ooservmg satellite ~.

monthly mean surface level humidity ^(Q) and provided three months data from 6July to 10 Oc-

376 _{INDIAN JRADIO} &SPACE PHYS, DECEMBER 1994

tober 1978. The SMMR was fl!lWll on Seasat spa- ::ecraft with the objective of collecting data of SST, WS, vertically integrated water vapour, cloud li- quid water content, and rain rate. The SMMR measured -microwav~ radiation at both vertical and horizontal polarizations at five frequencies (6.6, 10.7, 18.0, 21.0, and 37.0 GHz). The instrument is described elsewhere by Gloerson and Barath9 and Njoku eta/.lD The ph)'sical basis for converting th~

SMMR microwave measurements into geophysical units is described by Wilheit and Chang" and Pandey and Kakar12• In this paper, SMMR-derived ocean surface wind speed (WS), SST and ^W have been used along with Fleet Numerical Oceanogra- phic Centre (FNOC) model supplied sea surface pressure (SSP) to estimate LHF over the Indian Ocean.

3 Methodology

The evaporation can be determined by the bulk aerodynamic formula

... (1)

qa= 194.3 W- 84.69 W2

+

17.79 W3 -1.8 W4

+0.072 ^WS-151.59 '" (3)

The equation was derived using MONEX-79 ship observations. Hereinafter, we shall refer to ^qa as Q for the sake of standard practice in literature.

The six-hourly upper air observations collected from five USSR research vessels over the Arabian Sea and the Bay of Bengal in May-July 1979 were used. Eq. (3) was derived using about 600 observ- ations and it was validated using an independent set of about 200 observations. It estimates Q with Lm.S. error of 0.92 g/kg when atmospheric water vapour ranges between - 3 g/ cm2 and 7.5 g/cm2.

It may not give. reasonable results of Qfor values of W beyond the above range. It has been seen that Q estimated by Liu's global ^Q-W relation us- ing instantaneous data set yields Lm.S. error of 5.66 g/kg when compared with the observed ones. It shows that Liu's global Q-W relation is not valid over the tropical Indian oceanic region using instantaneous data set, which Hsu and Blanchard7 have also suggested.

L=597.0 - 0.6 SST

CD is the evaporation drag coefficient which is a function of wind speed (WS), given by Saunders'3 as

CD=A+Btanh (JWS2 -4.C )

where A= 1.65 x 10-3, B=8 X 10-5 and C=OA.

The density of air has been taken as 1.23 ~g/m3.

The specific humidity at sea surface is given by where ^qs and ^qa are the specific humidities at the sea surface and in the air near surface, and p is the air density in the air near surface. The latent heat flux can be obtained by multiplying E with la- tent heat of vaporization

L,

where

4 Results and discussion

It is clear from Eqs (1) and (2) that LHF calcu- lation requires the following parameters, SST, qs' qa' and WS. Most of these parameters are directly measurable by in situ observations on ships. When LHF calculated from the direct method (which uses ship observations of SST, WS and surface le- vel humidity) and that from the indirect method (which uses ship observations of SST and WS, and qa obtained from the Q-W relation of Gautam et a/.s based on W measurements from radiosonde observations instead of measured q^a) are com- pared, it has been found that the two methods agree within certain error bars. Figure 1 shows the scatter plot between LHF computed by the direct method and the indirect method. The' correlation coefficient between latent heat fluxes obtained by these two methods is 0.9, which is quite encourag- ing. It is clear that it is possible to obtain reason- able estimate of LHF using indirect method, which uses Q-W relation developed by Gautam ^{et a/.'/'.}

Seasat-SMMR-derived total precipitable water has also been used for the application of Q-W re- lation of Gautam et aL⁸ to derive surface level hu- midity. The parameter ^qa is then used to compute LHF, using SMMR-derived values of WS and SST.

Due to non-availability of coincident ship and sa- tellite observations, it has not been possible to val- idate the LHF derived by satellite data. However, since Seasat-SMMR dat" have been validated and the accuracies of different retrieved geophysical '" (2a)

... (2b) 0.622

qs=--es

Ps- es where

is the saturated vapour pressure at the sea surface, p= -4.928,5=23.55, R= -2937.0, Ps is the sea level pressure and Ts is sea _surface temperature.

Eqs 2(a) and (b) can also be used for calculating specific humidity near the surface, with T being interpreted as dew point temperature. The surface level humidity q^a is calculated by q^a-W relation given by Gautam et a/.sThe q^{a -}W relation is

I'I ^"I I

GAUTAM et aL: HEAT FWXES OVER THE INDIAN OCEAN 377

Fig. I-Scatter plot between latent heat fluxes (LHF) by the direct method and thelindirect method.

o 100 200 300 400 500

LHF BY DIRECT METHOD (Wlm2)

a

•

6

•

•

•

4

•

•

•

•

SST-SAT(K)

Seasat - SMMR OATA

••

•

•

2

•

• • •

• •••••

• ••••

· _{• •} ^:.

•

.!

•

•

•• ••

• • • • ^{• •}

•

o ...:r

-' o

JOO N

-€.

!

_z ²⁵⁰

Q~

olI(

-' 200

la.Ia:

.•..;

o

i: ¹⁵⁰

Fig. 2-Scatter plot between latent heat fluxes computed using Gautam's et aL8 Q-W relation and difference of sea surface

temperature (SST) and surface air temperature (SAT).

•

• •

• •

•

•

•

•

MONEX -79 DATA

. :

•• ^• •

: . ^., ,

.. , ^•

. ' ^•

• •

~

~ 200

a

is

z

>

CD

Il. 100.

::t~

o

400

....•

N

-l

_~

...•

o

^JOO

o

::t

~

LLJ~

!)OO

variables such as »: WS and SSThavebeen ex- tensively ~stablished using in situ observations during special campaigns, one can feel quite confi- dent that reasonable values' of LHF are obtained based on satellite data. Further, in order to con- firm our results. the variation of LHF with an indi- cator of atmospheric instability has been studied.

This indicator is defined as the difference of sea surface temperature and surface air temperature (SST - SAT). Greater the difference, greater will be the atmospheric instability, which will cause higher evaporation and thus will yield higher va- lues of LHF. Figure 2 shows the scatter diagram between LHF estimated by modifi.ed equation and SST - SAT. It shows reasonably good correlation between these two quantities (r= 0.65), which is a good qualitative indicator of the reliability of LHF estimated by our approach.

Liu's global monthly mean Q- W relation4 has also been used to estimate LHF using instantane- ous data set as suggested by Hsu and Blanchard7 and this lRF has been compared with LHF ob- tained using Gautarn's et aL8 relation using Sea- sat-SMMR data. Figure 3 shows the scatter plot between lRF by Gautarn's et aL Q- Wrelation and lliF by Liu's Q- W relation4• This Figure shows that Liu's ^Q-W relation overestirnates(

-100

WI

m2) the LHF as compared to estimates' obtained by Gautarn's ^{et aL8} relation. It is in agreement with the findings of Gautarn et aL8 that Liu's global Q-

N

... r

•

E ^.•..•.

• •

~

• •

z ••

• •

Q 300

•

~ -'

•

w

• •

a: ~ ²⁰⁰

_• ... :,

,-

^-'= ^mIL>-

^•••

^{:r-' 100}

^{• • I ••••, • ••••}

^{e ••••.••}

... ,

o

o 50 100 150 200 250 300

LHF BY GAUTAM'sel al~ RELATION (W/m 2 )

Fig. 3-Scatter plot between latent heat t1uxes computed using Gautam's et aP Q-W relation and Uu's Q-Wrelation.

W relation underestimates Q and hence overesti- mates LHF in the Indian Ocean. The variability of precipitable water in the tropics is very high.

Hence the monthly mean calculations of lliF are likely to be overestimated. Further the. data uti- lized here is of mid-August which specially charac- terizes the peak monsoon period with high values of moisture contents and very· strong surface winds. It gives rise to more specific humidity in the air and hence reduced lRF on instantaneous basis,

378 INDIAN J RADIO & SPACE PHYS, DECEMBER 1994

,

24

DEEP DEPRESSION CASE NORMAL CASE

LAND LAND

96 o80

20

92 96

84 88

LONGITUDE (-E)

o 80 20

16

I-

~

/ _}

_1f:iot

I -

¹⁶ _- ^-

z

12

r 7~ ~ 1~

^{12 ~2ao}

-

\'-'

;J

•..•

~

/ ^I

p

w

0

::>

I-

8~

w

-

^I-~

^{~ ,'0} 0 ~ ⁸

...J

~J~n I-

~

4

L.

'- ,

^...J

\

I

4

Fig. 4--Latent heat flux fields for deep depression case (13-17 August) over the Bay of Bengal.

Fig. 5-Latent heat flux fields for normal case (7-11 August) over the Bay of Bengal.

Finally, the LHF in deep depression (13-17 Au- gust) and normal case (7-11 August) have been compared. The LHF has been calculated using Seasat-SMMR data and then averaged over 5 days during deep depression and normal case over the Bay of Bengal. The LHF fields in these two cases have been studied along with SMMR-derived rain rates over the Bay of Bengal.

The detailed description of all the cyclonic storms, deep depression and low pressure systems for the year 1978 has been presented by Sriniva- san et al.14 Fortunately this deep depression has been fully covered by Seasat observations during its operation period of three months. The deep depression on 13 August was concentrated around 19SN and 89.0oEin the morning under the influ- ence of tropical storm from south China Sea. The deep depression was found to move slowly wes- tward and crossed south Orissa Coast on 15 Au- gust afternoon. It became weak after 17 August.

According to Indian Daily Weather Report (IDWR), published by India Meteorological De- partment, this system caused active to vigorous

monsoon in Orissa, coastal Andhra Pradesh, Tel- angana and east Madhya Pradesh.

Figures 4 and 5 show respectively the gradients of LHF over the Bay of Bengal in deep depres- sion and normal case. It can be seen from these fi- gures that LHF are found to be more ( - 200 WI

m2 in northern Bay of Bengal) in deep depression compared to in normal case (where it is - 120 WI

m2 in northern Bay of Bengal). The averaged SMMR-derived rain rates have also been seen over the same area (Figs 6 and 7). It is evident from these figures that higher rain rates in nor- thern Bay of Bengal are observed in deep depres- sion case compared to in normal case.

However, in this study it was impossible to crit- ically examine the deep depression and cyclonic storm because of unavailability of in situ data. The LHF shoud be studied more critically during deep depression and cyclonic storm over the Bay of Bengal along with in situ observations which would certainly enhance our understanding about the low pressure systems.

1'1 I ^{, 11'1 II It} ~ 1'1"1'1" I'!' I~'I UHli-IIIIIIII!1 IIIMI,1111nlUl>1 ill',j .,,,,,. ,

GAVTAM

,

et al.:HEAT FLUXES OVER THE INDIAN OCEAN 379

24

DEEP DEPRESSION CASE

24

NORMAL CA SE

LAND LAND

96 92

o

88

LONGITUDE (eE)

o .~

6

4

o

'80 20

12 16

LONGITUDE(·E ) 20

16

~\.( I ^{I .-- j}

^)~

f

^J

i-

0^•....•

z

i'2p_~~

0

lIJ :::>

l-I-

e:(

~ 8

...J

•...

e:(

...J

4

11 Fig. 6-SMMR-derived rain rate fields for deep depression case (13-17 August) over the Bay of Bengal.

Fig. 7-SMMR-derived rain rate fields -for normal case (7-11 August) over the Bay of Bengal.

5 Conclusions

This study demonstrates the application of new surface level humidity and precipitable water rela- tion for the estimation of l.HF using instantaneous data set. The comprehensive study of cyclone and deep depression using data from new generation microwave sensors like SSM/I along with in situ observation would be of immense help in enhanc- ing the knowledge about the complexity of low pressure systems over the Bay of Bengal. The heat fluxes thus estimated by satellite-borne passive microwave radiometry data would provide unique input for data assimilation in numerical models and air-sea interaction studies.

References

1 Pandey P C, Meteorol &Atmos Phys (Austria), 47 (1992) 165.

2 NjokuEG, ProclEEE(USA), 70 (1982) 728.

3 Liu W T, in Large-Scale Oceanographic Experiments and Satellites, edited by C Gautier and M Fienx (D Reidel, Hingham, Massachusetts, USA), 1984,205.

4 Liu WT, Mon Weather Re~( USA), 114 (198.) 1591.

5 Gairola R M, Gautam N&Pandey P C, IndUlnJ Radio &

Space Phys, 21 (1992) 143.

6 Liu WT,JGeophys Res(USA), 93 (1988)6749.

7 Hsu S A & Blanchard B W, J Geophys Res (USA), 94 (1989) 14539.

8 Gautam N, Basu S, Gairola R M Il. Pandey P C, Bound- Layer Meterol(Netherlands), 60 (1992) 179.

9 Gloersen P Il. Barath F T, IEEE J Ocean Eng (USA), OE-2 (1977) 172.

10 Njoku E G, Stacy J M Il. 8arath F T, IEEE J Ocean Eng (USA), OE-5 (1980) 100.

11 Wilheit T T & Chang ATe, Radio Sci (USA), 15 (1980) 525.

12 Pandey P C Ii. Kakar

It

K, IEEE Trans Geosci Remote Sens( USA), GE-21 (1983) 208.

13 Saunders P,^JMar Res(USA), 34 (2)(1976) 155.

14 Srinivasan V, Ramakrishnan A R &Jambunathan R, Mau- sam (India), 31 (1980) 495.